Dark Surfer-A Dark Pattern Analyzer & Classifier

Department of Computer Science and Engineering, Nalla Narasimha Reddy Education Society’s Group of Institutions, Hyderabad, India

Abstract

Dark Surfer is a project which is aimed to detect and combat dark patterns on websites using advanced language model [LLM]. The problem with dark patterns lies in their potential to undermine user trust, autonomy, and well-being. They can lead to frustration, confusion, and even financial harm for users who unwittingly fall victim to them. Furthermore, they erode the integrity of the digital ecosystem by prioritizing short-term gains for businesses over long-term relationships with customers. Dark patterns refer to design techniques used in user interfaces to manipulate users into taking actions they might not otherwise choose to take. These patterns often exploit psychological biases and can lead to unintended or undesirable outcomes for users. They can manifest in various ways, such as deceptive language, misleading visuals, hidden costs, or confusing interfaces. To avoid us cases we had come up with a website, which helps in analysing the dark patterns and detect them using URLS. We develop a classifier trained on a dataset of label UI elements, encompassing various types of dark patterns and benign design features. Leveraging natural language processing techniques and visual analysis, our classifier identifies deceptive design elements based on linguistic cues, visual attributes, and interaction patterns. So that we can add to the browser extension to analyse the number of dark patterns does the website contains, we also provided a flag site to detect the dark patterns, about the patterns using URL.